xxvn.] SOME PHOTOGRAPHS. 381 



Spectrum of manganese, showing the absorption of its triplet (at wave-length 

 about 4030) with its radiation. 



Spectrum of manganese, in which the triplet is again reversed. Here the 

 triplet, together with its two included bright lines, looks exactly like a group 

 of eight radiation lines, each reversed line giving the appearanc of two 

 bright lines. 



Photographs showing non-symmetrical absorption lines in the arc. 



Spectrum of silver, showing two lines at about wave-lengths 4054 '3 and 4210 '0. 

 Both lines are fluffy and reversed ; the less refrangible line is much more strongly 

 expanded on its more refrangible side, and is carried up to a much greater height 

 as a radiation line than its other side. The more refrangible line is more sym- 

 metrical, but presents the same phenomena to some extent, only in the opposite 

 direction, its less refrangible side being the most developed. 



Spectrum of rubidium, showing line at wave-length 4202. Here the two ends 

 of the line are produced by radiation alone, the central portion showing absorption 

 on its more refrangible side with fluffy shading on its less refrangible side. 



Spectra of strontium and calcium, showing the absorption of light due to 

 the poles. 



Photograph showing the " trumpeting " of absorption lines that is, 



the gradual widening of the absorption line as it 



recedes from a pole. 



Spectra of calcium, in which the reversal is seen to widen as we approach the 

 faint end produced by the cooler external region of the arc, thns showing absorp- 

 tion increasing with reduction of temperature. 



Spectrum of lead, showing that the lead line at wave-length 4058 also 

 trumpets. 



Spectrum showing the barium line at 4553 '4 trumpeting. Here the line, after 

 proceeding to a considerable distance from the hottest region of the arc as a fine 

 reversed line, gradually expands towards its extremity. 



Photographs of the spectrum of the flame of the arc. 



Flame spectrum of manganese, showing the reversal of the triplet in the 

 flame. 



Flame-spectra of calcium, showing the gradual extinction first of K then of H 

 as the flame recedes farthest from the arc. 



From these records of the phenomena we learn first, that not 

 all lines of a substance show reversal ; secondly, as already 

 stated, that those lines which are most easily reversed, that is to 

 say, which show the absorption most obviously in the arc itself, 

 are those which extend to the greatest distance in the flame; 

 and, thirdly, that we may have indications of absorption in 



